A temperature maintaining method for lithium titanate power battery suitable for low temperature environment

Abstract

The application discloses a lithium titanate power battery temperature maintaining method suitable for a low-temperature environment and particularly relates to the field of power battery thermal management control technology, and is used for solving the problem of low control reliability caused by the fuzzy responsibility coordination of a vehicle-mounted battery management system and a charging pile control system and the asynchronous state information in the existing battery preheating process; by acquiring battery temperature data and heater state data, a cause-effect correlation mode is identified based on historical coordinated operation data time sequence analysis when the battery temperature is lower than a threshold value, then the coordinated failure risk parameters of different preheating control right distribution schemes are predicted, the vulnerability of the optimal scheme is evaluated, the final execution scheme is selected according to the evaluation result, real-time data is analyzed to determine potential abnormal signs and their source types, and finally, the final scheme or a matched fault-tolerant control strategy is executed according to whether there are abnormal signs; the safety of the battery preheating process of the coordinated work of the double systems in the low-temperature environment is ensured.

Description

Technical Field

[0001] This invention relates to the field of power battery thermal management and control technology, and more specifically, to a method for maintaining the temperature of lithium titanate power batteries suitable for low-temperature environments. Background Technology

[0002] In new energy forklifts, especially those used in low-temperature environments such as cold chain warehousing, a common approach to ensure that the lithium titanate power battery can start, charge, and maintain sufficient power at low temperatures is to use off-site fixed charging piles to preheat the battery before charging. The charging piles provide external power, and the on-board battery management system monitors the battery temperature to jointly implement the heating operation, which constitutes a key link in the method of maintaining battery temperature in low-temperature environments. In existing technologies, the realization of this preheating process depends on data communication and command interaction between the on-board battery management system and the charging pile control system.

[0003] However, as two independent control units, the vehicle battery management system and the charging pile control system have inherent ambiguities in their functional boundaries and responsibilities. This leads to unclear responsibilities and inaccurate coordination during key operations such as system status judgment and anomaly handling decisions in the preheating process. At the same time, because they rely on intermittent periodic communication for status synchronization, they are prone to asynchronous, delayed, or failed status information under complex operating conditions or when communication is interfered with. This results in uncertainty and insufficient safety redundancy in the control logic of the entire preheating process, making it difficult to guarantee the control reliability and system safety under low temperature preheating conditions. This limits the effectiveness and universality of this temperature maintenance method in harsh industrial applications. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method for maintaining the temperature of lithium titanate power batteries suitable for low-temperature environments to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for maintaining the temperature of a lithium titanate power battery suitable for low-temperature environments includes:

[0007] S1. Obtain battery temperature data collected by the vehicle battery management system and heater status data collected by the charging pile control system;

[0008] S2. When the battery temperature data is lower than the preset preheating start temperature threshold, the battery temperature data and heater status data are analyzed in time series based on the historical collaborative operation data of the vehicle battery management system and the charging pile control system to identify causal relationship patterns.

[0009] S3. Based on the causal relationship model, predict the collaborative failure risk parameters corresponding to different preheating control allocation schemes;

[0010] S4. Based on the causal relationship model, conduct a causal counterfactual vulnerability assessment of the preheating control allocation scheme with the lowest collaborative failure risk parameter.

[0011] S5. Based on the results of the collaborative failure risk parameters and vulnerability assessment, select the final execution plan, and perform trend analysis on the real-time acquired battery temperature data and heater status data to determine potential abnormal signs and their source types.

[0012] S6. If there are no potential abnormal signs, execute the final execution plan; if there are potential abnormal signs, select the corresponding fault-tolerant control strategy according to the source type and execute the heating control.

[0013] Furthermore, S1 includes:

[0014] Through the communication interface between the vehicle battery management system and the charging pile control system, the battery temperature data periodically sent by the vehicle battery management system is received in real time.

[0015] Through the communication interface, the system receives real-time heater status data, including heater start / stop status and operating power, sent by the charging pile control system during heater operation.

[0016] Furthermore, S2 includes:

[0017] Extract time-series data segments with the same operating conditions as the current battery temperature data and heater status data from the historical collaborative operation data of the vehicle battery management system and the charging pile control system.

[0018] The battery temperature data sequence and heater status data sequence in the extracted time-series data segments are time-aligned to form a synchronized, collaborative operation data pair sequence;

[0019] The dynamic correlation strength between battery temperature data and heater status data in the synchronized collaborative operation data sequence under multiple time delays is quantified.

[0020] Based on the distribution and statistical significance of dynamic correlation strength under multiple time delays, a stable causal relationship from heater state data to battery temperature data is identified as a causal correlation pattern.

[0021] Furthermore, S3 includes:

[0022] Based on the causal influence strength and time delay from heater state data to battery temperature data in the causal association model, the causal chain maintenance conditions required to maintain effective heating are determined.

[0023] For each preheating control allocation scheme, analyze the control logic of the preheating control allocation scheme on the heater status data, and determine whether the control logic satisfies the causal chain maintenance condition.

[0024] The probability of a break in the causal relationship between heater state data and battery temperature data due to failure to meet the causal chain maintenance condition is quantified and used as a collaborative failure risk parameter corresponding to the preheating control allocation scheme.

[0025] Furthermore, S4 includes:

[0026] Based on the causal association model, at least one counterfactual condition is constructed that can disrupt the causal chain from heater state data to battery temperature data.

[0027] Under the premise of keeping other causal relationships unchanged in the causal relationship model, we infer the control effect of the preheating control allocation scheme with the lowest collaborative failure risk parameter on battery temperature data when counterfactual conditions occur.

[0028] Based on the degree of deviation between the simulated control effect and the expected control effect without counterfactual conditions, assess the vulnerability of the preheating control allocation scheme under abnormal conditions.

[0029] Furthermore, based on the degree of deviation between the simulated control effect and the expected control effect without counterfactual conditions, the vulnerability of the preheating control allocation scheme under abnormal conditions is assessed, including:

[0030] The control effect of the battery temperature data obtained under the counterfactual conditions is compared with the expected control effect under normal conditions based on the causal relationship model.

[0031] Calculate the difference between the two in terms of key performance indicators, where the key performance indicators include at least one of the decrease in temperature maintenance stability and the time delay in achieving the target temperature;

[0032] Based on the comparison between the degree of difference and the preset degree of difference threshold, the vulnerability level of the preheating control allocation scheme under abnormal conditions is determined.

[0033] Furthermore, S5 includes:

[0034] Based on the complementary relationship between the collaborative failure risk parameters and the vulnerability assessment results, a scheme selection rule is constructed, and the final execution scheme is selected from the schemes with the lowest collaborative failure risk parameters based on the scheme selection rule;

[0035] Extract the trend characteristics of change within the current time window from the real-time acquired battery temperature data and heater status data;

[0036] The matching degree is calculated between the changing trend characteristics of battery temperature data and the changing trend characteristics of heater status data to identify whether the changing trend characteristics of the two deviate from the normal cooperative pattern established by the causal relationship pattern.

[0037] When a deviation is identified, the deviation characteristics are compared with a pre-defined library of typical abnormal patterns to determine the source type of the potential abnormality.

[0038] Furthermore, the source types include those indicated by abnormal battery temperature data, those indicated by abnormal heater status data, and those indicated by abnormal coordination between the two.

[0039] Furthermore, S6 includes:

[0040] After identifying potential anomalies, the target fault-tolerant control strategy corresponding to the source type is matched from the preset set of fault-tolerant control strategies based on the determined source type of the potential anomalies.

[0041] When the source type is indicated by abnormal battery temperature data, the first type of fault-tolerant control strategy is executed;

[0042] When the source type is indicated by the abnormality of heater status data, the second type of fault-tolerant control strategy is executed;

[0043] When the source type is the type indicated by the abnormal relationship between the two, the third type of fault tolerance control strategy is executed.

[0044] Furthermore, the first type of fault-tolerant control strategy includes increasing the sampling frequency of battery temperature data and dynamically adjusting the control commands for heater status data based on the latest sampling data; the second type of fault-tolerant control strategy includes switching to a backup communication link to obtain heater status data and adjusting the control commands based on the redundancy data verification results; the third type of fault-tolerant control strategy includes activating a third-party temperature monitoring unit independent of the vehicle battery management system and the charging pile control system, and executing heating control based on the verification data of the third-party temperature monitoring unit.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. By establishing and utilizing the dynamic causal relationship model between the vehicle battery management system and the charging pile control system, a clear quantitative decision-making basis is provided for the coordinated control of the preheating process. This effectively overcomes the problem of unclear responsibilities caused by ambiguous functional boundaries. Based on the time-series analysis of historical coordinated operation data, the causal relationship model reflecting the real interaction law of the system is proactively identified. Based on this, the risk of coordinated failure of different control allocation schemes is proactively predicted and vulnerability is assessed. This makes the selection of the final execution scheme no longer a static division of authority, but a dynamic optimization process that integrates real-time risk quantification and anomaly resistance assessment. The originally ambiguous coordinated responsibility is transformed into a clear decision chain based on data and model-driven principles. This significantly improves the reliability and control accuracy of the dual-system coordinated operation under low-temperature preheating conditions, fundamentally avoiding control chaos or safety risks caused by inaccurate responsibilities, and ensuring the orderliness and controllability of the battery preheating process.

[0047] 2. By constructing a complete technical closed loop encompassing real-time trend analysis, anomaly prediction diagnosis, and directional fault-tolerant response, the system's state maintenance and autonomous recovery capabilities under complex operating conditions or communication disruptions are significantly enhanced. Through continuous monitoring of the coordinated change trends of battery temperature data and heater status data during preheating, and real-time comparison with identified causal relationship patterns, potential anomalies can be keenly detected and accurately located as originating from the battery side, heater side, or abnormal coordination relationship. Based on this, specific fault-tolerant control strategies strictly matching the anomaly source type are triggered, achieving automated closed-loop management from anomaly detection and root cause diagnosis to targeted handling. This makes the entire temperature maintenance system highly resilient to common interference factors such as intermittent communication lag and data anomalies, maximizing the continuity of the preheating function while ensuring battery safety. This improves the overall robustness and environmental adaptability of the lithium titanate power battery temperature maintenance method in low-temperature environments. Attached Figure Description

[0048] Figure 1 This is a flowchart of a method for maintaining the temperature of a lithium titanate power battery suitable for low-temperature environments according to the present invention. Detailed Implementation

[0049] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0050] Example: Figure 1 This invention provides a method for maintaining the temperature of a lithium titanate power battery suitable for low-temperature environments, comprising:

[0051] S1. Obtain battery temperature data collected by the vehicle battery management system and heater status data collected by the charging pile control system;

[0052] S2. When the battery temperature data is lower than the preset preheating start temperature threshold, the battery temperature data and heater status data are analyzed in time series based on the historical collaborative operation data of the vehicle battery management system and the charging pile control system to identify causal relationship patterns.

[0053] S3. Based on the causal relationship model, predict the collaborative failure risk parameters corresponding to different preheating control allocation schemes;

[0054] S4. Based on the causal relationship model, conduct a causal counterfactual vulnerability assessment of the preheating control allocation scheme with the lowest collaborative failure risk parameter.

[0055] S5. Based on the results of the collaborative failure risk parameters and vulnerability assessment, select the final execution plan, and perform trend analysis on the real-time acquired battery temperature data and heater status data to determine potential abnormal signs and their source types.

[0056] S6. If there are no potential abnormal signs, execute the final execution plan; if there are potential abnormal signs, select the corresponding fault-tolerant control strategy according to the source type and execute the heating control.

[0057] S1. Obtain battery temperature data collected by the vehicle battery management system and heater status data collected by the charging pile control system. Specifically, the implementation is as follows:

[0058] The on-board battery management system (BMS) and the charging pile control system are connected via a physical interface conforming to vehicle network communication standards, such as a controller area network (CAN) bus interface or an Ethernet interface. To achieve data exchange, both parties must pre-agree and configure the same communication protocol parameters, including the communication baud rate (e.g., 250 kbps or 500 kbps), data frame identifier format, and the order of data bytes. Battery temperature data is generated by the BMS through its internal periodic sampling program. The execution cycle of this periodic sampling program, for example, 100 milliseconds, is pre-set based on the real-time requirements of battery thermal management. When setting this cycle, a balance must be struck between data freshness and system communication load. If the cycle is too short, the communication load will be too high; if the cycle is too long, it may not be able to reflect rapid temperature changes in a timely manner. Therefore, a reasonable value is determined by evaluating the system's processing capacity and the fastest rate of temperature change. Before transmission, the battery temperature data is encapsulated by the BMS according to a predefined data frame format. The data frame contains at least one field representing the current temperature of the battery, in degrees Celsius. The temperature value must cover the battery's expected operating range in low-temperature environments, such as from -40 degrees Celsius to 80 degrees Celsius. The encapsulated data frame is continuously sent to the network node of the charging pile control system via the aforementioned communication interface at a set period.

[0059] Heater status data is actively generated and sent by the charging pile control system after the heater enters the operating state. The heater's operating state is determined by its internal control logic, typically triggered when the charging pile control system receives a preheating start command or detects that the battery temperature is below a certain internal setpoint. The heater status data is also encapsulated in a predefined, independent data frame format. This data frame contains a Boolean field indicating whether the heater is currently in a start or stop state, and a numerical field representing the heater's current operating power in kilowatts. The transmission of this data frame is not strictly periodic; it is tied to the heater's operating state. During the heater's start-up state, the charging pile control system sends the current heater status data through the same communication interface at a certain update interval, such as 500 milliseconds. This update interval is set considering the minimum response time of the heater power adjustment command and communication efficiency to ensure that status updates keep pace with control dynamics.

[0060] On the data receiving side, a receiving unit listens for and captures data through the same communication interface. The receiving unit continuously monitors the communication bus, and triggers a data reading operation when the detected data frame identifier matches a pre-configured target identifier. The acquired raw data byte stream is parsed according to a pre-agreed data frame format. For battery temperature data frames, binary data representing the temperature value is extracted from a specified byte position and converted into a floating-point number in degrees Celsius, for example, by multiplying the binary two's complement value by a coefficient of 0.1 to convert it into the actual temperature value. For heater status data frames, Boolean values ​​representing start / stop status and numerical values ​​representing operating power are extracted from different byte positions within the frame. After successful parsing, both battery temperature data and heater status data are temporarily stored as timestamped data points. The timestamp records the system time when the data was successfully parsed, with the unit of the timestamp being milliseconds. The timestamped data sequence constitutes the basic data source for subsequent steps of real-time analysis or historical backtracking. By providing a detailed description of the physical connection, protocol parameter settings, data frame structure definition, transmission cycle and trigger logic settings, parsing and conversion methods, and data storage format, the method ensures that battery temperature data and heater status data can be reliably, in real time, and in a uniform format, providing accurate and traceable data input for the entire method.

[0061] S2. When the battery temperature data is lower than the preset preheating start-up temperature threshold, a time-series analysis is performed on the battery temperature data and heater status data based on the historical collaborative operation data of the vehicle battery management system and the charging pile control system to identify causal relationship patterns. Specifically, the implementation is as follows:

[0062] The trigger condition for this step is that the real-time monitored battery temperature data falls below a preset preheating start-up temperature threshold. The preheating start-up temperature threshold is a pre-set fixed value, such as 5 degrees Celsius above zero. This threshold is based on the technical characteristics of lithium titanate batteries; when the battery temperature falls below this value, its charging and discharging performance and safety will significantly decrease, thus requiring the preheating process to be initiated. The specific method for determining whether a trigger has been triggered is to compare the real-time acquired battery temperature data with the preheating start-up temperature threshold. If the battery temperature data is lower than the preheating start-up temperature threshold, the trigger condition is met. Once this trigger condition is met, the process of retrieving and analyzing historical collaborative operation data is initiated. The historical collaborative operation data referred to here is a collection of battery temperature data sequences and heater status data sequences with precise timestamps, continuously acquired and stored by the method described in step S1 over multiple complete preheating cycles.

[0063] First, time-series data segments matching the currently monitored operating conditions are extracted from the stored historical collaborative operation database. The current operating conditions are primarily defined by the numerical range and rate of change of the real-time acquired battery temperature data, as well as the start-stop pattern of the heater status data. For example, the current battery temperature is between -15°C and +3°C with a gradual change, while the heater is operating at a fixed duty cycle. The extraction process involves iterating through and comparing data in the historical database using a sliding window of a preset time length. This sliding window length, for example, 30 minutes, is set based on the average time taken for a typical complete process from start-up preheating to the battery reaching a rechargeable state. Specifically, segments in the historical data where the numerical range and change pattern of both battery temperature data and heater status data are highly similar to the current characteristics within a continuous time period are identified and extracted as candidate time-series data segments. To ensure statistical significance for subsequent analysis, multiple, such as five or more, non-overlapping segments are typically extracted.

[0064] Next, the battery temperature data sequence and heater status data sequence in each extracted time-series data segment are precisely time-aligned to eliminate minor time base deviations that may exist during the original data acquisition and to form a strictly synchronized, collaboratively running data pair sequence. The alignment operation uses a more precise time source as a reference, such as the timestamp sequence corresponding to the battery temperature data acquisition point as the reference time axis. For each reference time point, a linear interpolation algorithm is used to estimate the corresponding heater status data value. Specifically, for a given reference time point, in the timestamp sequence of the heater status data, the two data points immediately before and immediately after it are found. Then, assuming that the change in state value between these two known points is uniform and linear, the estimated heater status value at the reference time point is calculated proportionally to the time. After this alignment process, each battery temperature data point is precisely paired with a calculated heater status data point, and all the paired data arranged in chronological order constitute a synchronized, collaboratively running data pair sequence.

[0065] Then, for each synchronized, collaboratively running data pair sequence, a quantitative calculation of the dynamic correlation strength is performed to assess the degree of correlation between changes in heater state data and battery temperature data after experiencing different time delays. The time delay range examined here needs to cover a reasonable interval from control action to the generation of observable thermal effects, for example, from 10 seconds to 300 seconds. The lower limit of this range, for example, 10 seconds, considers the shortest time it takes for a change in heater power to be detected by the sensor; the upper limit, for example, 300 seconds, considers the longest possible time for heat to transfer from the heating element to the battery core and cause a significant temperature change. During the calculation, a series of candidate time delay values ​​with equal or non-equal intervals are selected within this range, such as 30 seconds, 60 seconds, 120 seconds, and 180 seconds. For each candidate time delay value, the heater state data sequence in the entire collaboratively running data pair sequence is shifted forward along the time axis by that delay amount. After the shift, the Pearson product-moment correlation coefficient between the shifted heater state data sequence and the original battery temperature data sequence is calculated. Calculating the Pearson product-moment correlation coefficient requires both sequences to be of equal length, with a value ranging from -1 to +1. The result is a dimensionless numerical value that measures the degree of linear correlation between the two sequences. By repeating this calculation for all candidate time delay values, a curve or a list of values ​​describing how the dynamic correlation strength changes with the delay time can be obtained.

[0066] Finally, based on the calculated distribution of dynamic correlation strength under multiple time delays, stable and significant causal relationships are identified and formalized into causal association patterns. The identification process first requires a dynamic correlation strength significance threshold, for example, 0.6. This threshold is determined based on statistical analysis of correlation strength values ​​calculated from a large amount of historical normal operating data; for example, the upper quartile of the historical correlation strength value distribution can be used as the threshold. During identification, the absolute value of the dynamic correlation strength calculated for each candidate time delay is compared with this threshold. Only delay values ​​with an absolute value greater than this threshold are considered to have a significant correlation. Then, within the subset of significant delay values ​​initially selected through comparison, a statistical significance test is performed, calculating the statistical significance test P-value corresponding to the correlation coefficient for each delay value. Each statistical significance test P-value is compared with a preset statistical significance level threshold, for example, 0.05. A p-value less than the statistical significance threshold indicates that the correlation is less than 5% likely to be caused by random factors, thus being statistically significant. Finally, from among the lag values ​​that satisfy both the absolute value of the dynamic association strength being greater than the significance threshold and the p-value being less than the statistical significance threshold, the time lag with the largest absolute value of the dynamic association strength is selected as the principal lag time of the causal effect. The finally identified causal relationship pattern will be explicitly described as structured knowledge containing the following key parameters: causal direction, average lag time representing the speed of influence transmission, association strength coefficient representing the strength of the influence, and the statistical confidence level of the relationship. This pattern will serve as the quantitative basis for subsequent prediction and evaluation steps. If, for a given time segment, the absolute value of the dynamic association strength under all candidate lags does not exceed the significance threshold, or if it does exceed the threshold but the p-value is greater than the statistical significance threshold, then that segment is considered an invalid sample and not used for the final pattern induction; the final pattern is formed based on conclusions supported by a majority of valid samples.

[0067] S3. Based on the causal relationship model, predict the synergistic failure risk parameters corresponding to different preheating control allocation schemes. The specific implementation is as follows:

[0068] The input to this step is the causal relationship pattern identified and output in step S2. This pattern explicitly includes two core quantitative parameters: the causal influence strength and the time delay from the heater status data to the battery temperature data. For example, an identified causal relationship pattern might be specifically described as having a causal influence strength of 0.85 and a time delay of 15 seconds. The goal of the prediction is to calculate the corresponding collaborative failure risk parameters for multiple different preheating control allocation schemes. This parameter is a numerical index used to quantify the probability of failure of the scheme during implementation, and its value range is typically defined between 0 and 1.

[0069] First, based on the specific causal influence strength and time delay in the causal association model, the causal chain maintenance conditions necessary to maintain effective heating are determined. The causal chain maintenance conditions are a set of rules constraining the dynamic relationship between heater state data and battery temperature data, aiming to ensure that identified causal relationships are maintained during actual control. Specifically, the causal chain maintenance conditions include at least two sub-conditions. The first sub-condition is the response consistency condition, which requires that the heater state data's response to control commands must be completed within a specific time window. The upper limit of this time window is directly derived from the time delay parameter in the causal association model. For example, a time delay of 15 seconds requires that the duration of the change in heater state data from receiving the command to reaching the expected state in actual control should be less than 15 seconds. The second sub-condition is the influence sufficiency condition, which requires that the magnitude of change in heater state data be sufficient to trigger the expected change in battery temperature data. This condition is quantified by a preset influence sufficiency threshold, which is calculated based on the causal influence strength parameter in the causal association model. The calculation method involves multiplying the causal influence strength by a preset scaling factor, which is a positive number less than 1, such as 0.9, to provide a margin for performance fluctuations in actual operation. For example, if the causal influence strength is 0.85, multiplying it by the scaling factor 0.9 yields an influence sufficiency threshold of 0.765. This means that in actual control, the standardized correlation between changes in heater state data and the resulting changes in battery temperature data should be no less than 0.765.

[0070] Secondly, for each preheating control allocation scheme to be evaluated, the control logic of the scheme on the heater state data is analyzed, and it is determined whether this control logic satisfies the aforementioned causal chain maintenance condition. The preheating control allocation scheme defines the authority, timing, and logical rules for the on-board battery management system and the charging pile control system to issue control commands to the heater during the preheating process. For example, one scheme may have the on-board battery management system having exclusive control, while another scheme may have the charging pile control system making autonomous decisions based on battery temperature data. When analyzing the control logic, it is necessary to examine the control command generation rules, command transmission paths, and expected heater response characteristics specified in the scheme. The process of determining whether the causal chain maintenance condition is satisfied is performed item by item. For the response consistency condition, it is necessary to evaluate the estimated time from the issuance of the command to the expected change in the heater state data under the control logic specified in the scheme, and compare this estimated time with the upper limit of the time window specified in the causal chain maintenance condition. If the estimated time is less than the upper limit, it is determined that this sub-condition is satisfied. For the sufficiency condition, it is necessary to simulate the change pattern of heater state data under the scheme control logic based on historical data or system models, and estimate the possible correlation between this change pattern and the change in battery temperature data. This estimated correlation is compared with the sufficiency threshold of influence specified in the causal chain maintenance condition. If the estimated correlation is greater than or equal to the sufficiency threshold of influence, then this sub-condition is determined to be satisfied. Only when the control logic simultaneously satisfies all causal chain maintenance sub-conditions is the preheating control allocation scheme determined to satisfy the causal chain maintenance condition as a whole.

[0071] Finally, for each preheating control allocation scheme, the probability of a break in the causal relationship between heater state data and battery temperature data due to failure to meet the causal chain maintenance condition is quantified, and this probability is used as the corresponding co-failure risk parameter for that scheme. The probability of breakage is calculated using a multi-factor weighted evaluation model. The input to this model is the binary judgment result of whether each causal chain maintenance sub-condition is met in the previous step, and a deviation quantification value representing the degree of non-metdency. For example, if the response consistency condition is not met, the deviation can be quantified as the percentage of the estimated time exceeding the upper limit of the time window; if the influence sufficiency condition is not met, the deviation can be quantified as the difference between the influence sufficiency threshold and the estimated correlation. When calculating the co-failure risk parameter, a weight factor is assigned to each sub-condition, which reflects the importance of the sub-condition in maintaining the causal relationship, and the sum of the weight factors is 1. The specific values ​​of the weighting factors can be set based on the analysis of historical failure cases or expert experience. For example, if historical data analysis reveals that failures due to response delay account for a higher proportion, the weighting factor for the response consistency condition can be set to 0.6, and the weighting factor for the sufficiency condition can be set to 0.4 accordingly. The specific calculation method for the co-failure risk parameter is as follows: First, check whether each sub-condition is satisfied; for satisfied sub-conditions, their contribution value is 0; for unsatisfied sub-conditions, multiply their calculated deviation quantification value by their corresponding weighting factor to obtain the risk contribution value of that sub-condition; then, sum the risk contribution values ​​of all sub-conditions to obtain the preliminary risk value; finally, to avoid the preliminary risk value exceeding the defined domain, it can be mapped to the range of 0 to 1 using a normalization function, such as an sigmoid function, and the final output is the co-failure risk parameter corresponding to the preheating control allocation scheme. If a scheme fully satisfies all causal chain maintenance conditions, its co-failure risk parameter is 0. In this way, each preheating control allocation scheme is assigned a quantitative collaborative failure risk parameter. The value of this parameter directly reflects the risk of the key causal relationship identified in step S2 breaking under the control logic of the scheme, thus providing a clear quantitative basis for subsequent decision-making.

[0072] S4. Based on the causal relationship model, a causal counterfactual vulnerability assessment is conducted on the preheating control allocation scheme with the lowest collaborative failure risk parameter. The specific implementation is as follows:

[0073] The inputs to this step include the causal relationship patterns identified in step S2, and the co-failure risk parameters calculated in step S3 for all candidate preheating control allocation schemes. The evaluation focuses on the preheating control allocation scheme with the lowest co-failure risk parameter value, which is considered the scheme with the lowest conventional risk. The vulnerability assessment aims to reveal the robustness of this scheme when encountering unconventional and anomalous conditions; the output of the assessment is a qualitative or quantitative conclusion characterizing its vulnerability level.

[0074] First, based on the causal association model, at least one counterfactual condition is constructed that disrupts the causal chain from heater state data to battery temperature data. A counterfactual condition is a hypothetical system state or external event that deviates from normal expectations, designed to proactively challenge the stability of identified causal relationships. The construction process closely relies on the quantified time delay parameter and causal influence strength parameter in the causal association model. For the time delay dimension of the counterfactual condition, the construction logic assumes a delay in communication or execution between systems, specifically implemented by defining a delay multiplier. For example, setting the delay multiplier to 2 means that the time delay under the counterfactual condition is calculated as the normal time delay parameter value in the causal association model multiplied by 2. For the causal influence strength dimension of the counterfactual condition, the construction logic assumes a decrease in heat transfer efficiency, specifically implemented by defining an intensity attenuation coefficient. For example, if the intensity attenuation coefficient is set to 0.5, then the causal influence intensity under counterfactual conditions is calculated as the normal causal influence intensity parameter value in the causal association model multiplied by 0.5. 0.5 is a multiplier or coefficient value, which is preset based on the analysis of potential failure modes. For example, a communication delay multiplication factor of 2 corresponds to the typical delay increment under a single-channel failure.

[0075] Secondly, while maintaining the causal relationships in the causal relationship model that are not directly altered by the counterfactual condition, the control effect of the preheating control allocation scheme with the lowest selected collaborative failure risk parameter on the battery temperature data is deduced when the constructed counterfactual condition occurs. The deduction is a simulation analysis process based on logic and causal models. During the deduction, the predetermined control logic of the preheating control allocation scheme is used as the simulation input. Then, the selected counterfactual condition is introduced as a mandatory constraint into the simulation environment. This means that during the simulation, the model parameters used to calculate the changes in heater state data or their impact on battery temperature data are temporarily replaced with new values ​​that conform to the definition of the counterfactual condition. For example, when a counterfactual condition for time delay is introduced, the inherent processing delay parameter in the submodule simulating control command transmission and execution is replaced with an increased time delay calculated using the aforementioned method. Subsequently, according to the control logic of the scheme, starting from initial battery temperature data and heater state data, the decision-making process of each control cycle is simulated step by step. In each cycle, control commands are generated based on the current simulated battery temperature data and the scheme logic. Then, combined with counterfactual constraints, the expected heater state data under this command is calculated. Finally, using the basic influence relationship model from heater state data to battery temperature data in the causal relationship model, but replacing the time delay parameter and influence intensity parameter with their corresponding values ​​under counterfactual conditions, the change in battery temperature data simulated in this cycle is calculated. This process is iterated until the simulation time reaches the preheating set duration, thus obtaining a sequence of battery temperature data for the entire simulation period. This sequence represents the derived control effect.

[0076] Next, based on the deduced control effect, its deviation from the expected control effect without counterfactual conditions is evaluated. This process first requires establishing the expected control effect without counterfactual conditions. This expected effect is the battery temperature data sequence simulated using the same preheating control allocation scheme control logic and normal causal relationship mode parameters, through the aforementioned deduction method. Then, the deduced effect sequence under counterfactual conditions is compared with the normal expected effect sequence. The comparison is achieved by calculating the difference between the two on predefined key performance indicators. Key performance indicators include at least one of the following: the decrease in temperature maintenance stability and the time delay in achieving the target temperature. The decrease in temperature maintenance stability is calculated by calculating the standard deviation of the battery temperature data in the counterfactual deduced effect sequence and the standard deviation of the battery temperature data in the normal expected effect sequence within a fixed time window after the preheating process enters steady state, for example, the last 60 seconds. Then, the standard deviation under the counterfactual conditions is subtracted from the standard deviation under the normal conditions, and the difference is the decrease in temperature maintenance stability, expressed in degrees Celsius. The method for calculating the time delay of achieving the target temperature is to find the moment when the battery temperature data first reaches or exceeds the preset target temperature value, such as 15 degrees Celsius above zero, from the counterfactual effect sequence and the normal expected effect sequence, respectively, and calculate the time difference between these two moments. This time difference is the time delay of achieving the target temperature, and its unit is seconds.

[0077] Finally, based on the comparison between the calculated difference and the preset difference threshold, the vulnerability level of the preheating control allocation scheme under abnormal conditions is determined. The preset difference threshold is a threshold value used to classify different vulnerability levels and needs to be set separately for each key performance indicator. For example, for the target temperature achievement time delay, the first-level preset difference threshold is set to 10% of the normal expected achievement time, and the second-level preset difference threshold is set to 50% of the normal expected achievement time. The first-level preset difference threshold, for example, corresponds to a 60-second delay, and its setting is based on considering the allowable preheating completion time deviation range under normal operating fluctuations. The second-level preset difference threshold, for example, corresponds to a 300-second delay, and its setting is based on the assumption that exceeding this delay will severely hinder the charging process, and the preheating function will be considered essentially failed. During the evaluation, the calculated target temperature achievement time delay is compared sequentially with the first-level and second-level preset difference thresholds. If the delay is less than the first-level preset difference threshold, the vulnerability level is determined to be low; if the delay is greater than or equal to the first-level preset difference threshold but less than the second-level preset difference threshold, the vulnerability level is determined to be medium; if the delay is greater than or equal to the second-level preset difference threshold, the vulnerability level is determined to be high. The assessment of the decrease in temperature stability follows the same logic, but uses its corresponding preset difference threshold in degrees Celsius for comparison. For the preheating control allocation scheme being evaluated, its final comprehensive vulnerability level is the highest among all the levels determined by the evaluated key performance indicators. For example, if the time delay is determined to be medium and the decrease in temperature stability is determined to be low, the final vulnerability level is medium.

[0078] S5. Based on the results of the collaborative failure risk parameters and vulnerability assessment, the final implementation plan is selected, and trend analysis is performed on the real-time acquired battery temperature data and heater status data to determine potential anomalies and their source types. The specific implementation is as follows:

[0079] The inputs to this step include the co-failure risk parameters for each preheating control allocation scheme calculated via step S3, and the vulnerability level of the scheme with the lowest co-failure risk parameters evaluated via step S4. Simultaneously, this step requires continuous access to the real-time battery temperature data and heater status data streams described in step S1, and relies on the causal relationship patterns identified in step S2.

[0080] First, a scheme selection rule is constructed based on the complementary relationship between the collaborative failure risk parameters and the vulnerability assessment results. Based on this rule, the final execution scheme is selected from the schemes with the lowest collaborative failure risk parameters. The scheme selection rule is a set of decision-making logic designed to consider both the conventional risks and abnormal vulnerabilities of a scheme. This rule maps vulnerability levels to a numerical vulnerability adjustment coefficient; for example, low-level vulnerability is mapped to a coefficient of 1.0, medium-level vulnerability to a coefficient of 1.5, and high-level vulnerability to a coefficient of 2.0. The vulnerability adjustment coefficient is set based on the fact that different vulnerability levels correspond to different degrees of performance degradation of the scheme under abnormal conditions. This coefficient value is an estimate of the proportion of possible additional performance loss; for example, a medium-level vulnerability is estimated to lead to an additional 50% performance risk, hence the coefficient is 1.5. Next, a decision index is constructed. The decision index is calculated by multiplying the original collaborative failure risk parameter value of the scheme with the lowest collaborative failure risk parameter by the corresponding vulnerability adjustment coefficient. The calculated decision index value is used for the final judgment. Simultaneously, a scheme adoption decision threshold needs to be set, for example, 0.35. The decision threshold for scheme adoption is set based on statistical analysis of historical successful operation cases. For example, in all successful preheating processes in history, the decision index value after comprehensive evaluation was below 0.35, so this value is used as the safety boundary. The specific rule is as follows: if the calculated decision index value is less than the scheme adoption decision threshold, the scheme with the lowest collaborative failure risk parameter is determined to be directly adopted as the final execution scheme; if the decision index value is greater than or equal to the scheme adoption decision threshold, it means that the comprehensive risk brought by its abnormal vulnerability is too high. At this time, the scheme will be abandoned, and the same rule will be applied again to the scheme with the second lowest collaborative failure risk parameter. That is, its decision index value will be calculated and compared with the scheme adoption decision threshold until a scheme with a decision index value less than the scheme adoption decision threshold is selected as the final execution scheme.

[0081] Secondly, while the final execution plan begins implementing preheating control, trend analysis is performed in parallel on the real-time acquired battery temperature data and heater status data. The analysis first requires extracting the trend characteristics within the current time window. The current time window refers to a continuous period of time tracing back from the current moment; its length, for example, 180 seconds, should be sufficient to cover the complete causal cycle from heater activation to observable changes in battery temperature. For the battery temperature data sequence within this time window, the extracted trend characteristics include the linear fitting slope of the sequence, obtained by fitting a straight line using the least squares method; the slope is in degrees Celsius per second, representing the average rate of temperature change; and the standard deviation of the sequence, also in degrees Celsius, representing the magnitude of temperature fluctuation during this period. For the heater status data sequence within the same time window, which is time-aligned with the battery temperature data, the extracted trend characteristics include the cumulative duration of the heater's start-up state (duty cycle), a dimensionless value between 0 and 1; and the average heater operating power, in kilowatts. These features together constitute a quantitative feature vector describing the current instantaneous operating state of the system.

[0082] Then, the matching degree of the extracted battery temperature data trend characteristics and the heater state data trend characteristics is calculated to identify whether they deviate from the normal cooperative pattern established by the causal association model. The normal cooperative pattern is derived from the causal association model identified in step S2, and it specifically defines the quantitative relationship that should be satisfied between the changing trend characteristics of the heater state data and the changing trend characteristics of the battery temperature data when effective heating occurs. For example, a normal cooperative pattern may stipulate that when the heater duty cycle is greater than 0.5 and the average power is greater than 5 kW, the linear fitting slope of the battery temperature data should be greater than 0.02 degrees Celsius per second. The matching degree calculation is specifically implemented through a conditional verification function, which takes the extracted real-time feature vector as input and checks whether each feature component satisfies the corresponding constraint conditions defined in the normal cooperative pattern. If all conditions are satisfied, the matching degree calculation result is passed, indicating no deviation. If at least one condition is not satisfied, the matching degree calculation result is failed, and the specific feature condition(s) that were not satisfied, as well as the numerical difference between the actual feature value and the pattern requirement value, is recorded. This difference is quantified as the deviation feature quantity.

[0083] Finally, when the matching degree calculation identifies a deviation, the recorded deviation features are compared with a pre-defined library of typical anomaly patterns to determine the source type of the potential anomaly. The library of typical anomaly patterns is pre-established through analysis of historical fault data or simulation of abnormal scenarios. Each typical anomaly pattern record in the library contains a feature deviation fingerprint vector and a corresponding source type label. The feature deviation fingerprint vector records the typical deviation direction and magnitude of the changing trend features of battery temperature data and heater status data relative to the normal cooperative pattern when this type of anomaly occurs. For example, a source type indicated by anomaly in battery temperature data might show a fingerprint vector indicating a significant negative deviation in the linear fitting slope of the battery temperature data, while the deviation of various features in the heater status data is close to zero. The comparison process calculates the similarity between the deviation feature vector obtained from the current real-time analysis and the feature deviation fingerprint vector of each typical anomaly pattern in the library. The similarity can be calculated using vector cosine similarity, with a value range between -1 and +1; a larger value indicates a greater similarity between the two patterns. The typical anomaly pattern with the highest cosine similarity is selected, and its corresponding source type is determined as the source type of the current potential anomaly. If all calculated cosine similarities are below a preset anomaly pattern matching threshold, such as 0.7, the deviation is marked as an unknown type of anomaly. The anomaly pattern matching threshold of 0.7 is set to ensure the reliability of the matching, based on the statistical results of the similarity distribution between a large number of known type anomaly samples and erroneous deviation samples. Through the above process, during the final execution of the plan, continuous monitoring of the real-time running status and early anomaly diagnosis are achieved, providing key input for the decision in step S6.

[0084] S6. If no potential anomalies are detected, the final execution plan will be executed; if potential anomalies are detected, the corresponding fault-tolerant control strategy will be selected based on the source type to execute the heating control, specifically as follows:

[0085] The input to this step is the judgment result analyzed and output in real time by step S5. This result includes two possible states: one is that no potential abnormal signs are found, and the other is that potential abnormal signs are found, along with the source type of the identified potential abnormal signs. The logical branches and execution actions of this step are all based on this input state.

[0086] First, after identifying potential anomalies, based on the source type of the potential anomalies determined in step S5, a target fault-tolerant control strategy corresponding to the source type is matched from a pre-defined set of fault-tolerant control strategies. The pre-defined set of fault-tolerant control strategies is a predefined data structure, such as a lookup table, which is pre-built before deployment by analyzing historical failure cases and expert experience. Each record in the table explicitly associates a specific source type of a potential anomaly with a specific fault-tolerant control strategy identifier. The matching process is a direct table lookup operation: using the source type string output in step S5 as the query key, a search is performed in the pre-defined set of fault-tolerant control strategies. The record that perfectly matches the source type field is found, and the specific fault-tolerant control strategy identifier pointed to by this record is selected as the target fault-tolerant control strategy, and the specific strategy logic represented by that identifier is ready to be executed.

[0087] When the source type is indicated by abnormal battery temperature data, the first type of fault-tolerant control strategy is executed. The core idea of ​​this strategy is to enhance the real-time performance of temperature sensing and the agility of control commands to cope with possible sensor reading anomalies or data transmission delays. Specific implementation includes two main actions. The first action is to increase the sampling frequency of battery temperature data, for example, by sending a reconfiguration command to the vehicle battery management system to shorten its periodic sampling interval for the battery temperature sensor from once every 100 milliseconds to once every 50 milliseconds. This new 50-millisecond sampling interval is set to ensure that the data update cycle is shorter than the characteristic time of rapid temperature changes that may occur during battery heating, while not exceeding the minimum sampling interval supported by the vehicle battery management system's data acquisition unit. The second action is to dynamically adjust the control commands for heater status data based on the latest sampled data. Dynamic adjustment means that the calculation logic of the control commands switches from relying on historical data over a longer time window to being based on data within a significantly shortened recent sliding time window, for example, using only the most recent 10 sampling points, i.e., the battery temperature data sequence within the most recent 500 milliseconds. First, the instantaneous rate of change of the short sequence is calculated, which is the difference between the last sampled value and the first sampled value divided by the time span. Then, the basic control command generated by the final execution scheme is modified according to a preset dynamic adjustment rule. This rule, for example, stipulates that if the calculated instantaneous rate of change is less than 50% of the expected rate of change benchmark value obtained from historical normal data statistics, the target power output value of the heater in the current control command will be increased by 20% based on the value calculated by the final execution scheme. The 20% adjustment is set to provide sufficient compensation without overreacting. The adjusted control command is immediately sent to the execution unit.

[0088] When the source type is indicated by an anomaly in the heater status data, the second type of fault-tolerant control strategy is executed. The core idea of ​​this strategy is to avoid the primary data link, which may be faulty or subject to interference, and to perform cross-validation and restore reliable control through a backup data source. The implementation includes two main actions. The first action is to switch to the backup communication link to obtain heater status data. The backup communication link refers to a pre-laid physical or logical communication path independent of the current primary communication channel, such as a pre-deployed pair of redundant differential signal lines in the controller area network bus architecture. The switching operation is achieved by sending a digital switching command to a communication gateway or multiplexer, which switches the heater status data source input port from the port connected to the primary link to the port connected to the backup link. The second action is to adjust the control command based on the redundancy data verification results. Redundancy data verification refers to simultaneously obtaining a physical quantity that indirectly reflects the heater's operating status from an independent data source, such as reading the real-time current value from a Hall current sensor connected in series in the heater power supply circuit. Using a known relationship model between heater power and current, an equivalent heating power is estimated based on the current value. Next, the operating power field value from the heater status data newly acquired from the backup communication link is compared with the equivalent power value estimated by the current, and the percentage of the absolute value of the difference between the two relative to the current estimate is calculated. This percentage value is then compared with a preset verification difference threshold, such as 15%. This 15% threshold takes into account the accuracy of the current sensor, model error, and the allowable system fluctuation range. If the difference percentage is less than 15%, the backup link data is considered reliable, and control commands will continue to be generated based on this backup link data. If the difference percentage is greater than or equal to 15%, it is determined that the backup link data may also be abnormal. In this case, the control commands will no longer rely on the heater status data, but will instead be based on the equivalent power value estimated by the current, and a preset conservative power control model will be used, for example, using 80% of the current estimate power value as the command output, to leave a safety margin and prevent overheating.

[0089] When the source type is indicated by an abnormality in the coordination relationship between the two systems, a third-type fault-tolerant control strategy is implemented. The core idea of ​​this strategy is to introduce a completely independent third-party data source as a control benchmark when the data coordination relationship between the vehicle battery management system and the charging pile control system becomes disordered, ensuring basic safe heating. The implementation involves two main actions. The first action is to activate a third-party temperature monitoring unit independent of both the vehicle battery management system and the charging pile control system. This unit is a physically independent temperature sensing and data reporting device. Its sensor probes are attached to key temperature measurement points on the battery pack, and it has its own power supply and dedicated wireless communication module, such as using the Zigbee protocol. Activation means sending an activation command to the unit via a wireless network, causing it to enter working mode from a low-power sleep mode and begin collecting battery temperature data at a set interval, such as every 2 seconds, and transmitting the data through an independent wireless network channel. The second action is to execute heating control based on the verification data from the third-party temperature monitoring unit. This involves continuously receiving and parsing the third-party temperature data. The control logic will primarily operate based on this third-party temperature data, establishing a simplified independent control loop. For example, a preset safe heating target temperature, such as 10 degrees Celsius above zero, can be used. The control rule is as follows: when the received third-party temperature data is below 10 degrees Celsius, a fixed safe power heating command is continuously sent to the heater. This power value, for example, 60% of the rated power, is a preset safe value. When the third-party temperature data reaches or exceeds 10 degrees Celsius, a stop heating command is sent. This method aims to completely eliminate the reliance on potentially distorted data coordination relationships, relying on an independent and reliable data source to implement a simplified yet safe temperature control logic, prioritizing the prevention of battery overheating or overcooling.

[0090] If step S5 determines that there are no potential anomalies, no fault-tolerant control strategy branch is triggered; instead, the final execution plan selected in step S5 is executed directly. Executing the final execution plan means continuously sending control commands to the heater according to the preheating control allocation logic, control command generation algorithm, and communication protocol specified in the plan, and performing closed-loop control based on normal battery temperature data and heater status data feedback until the preheating process is completed or the termination condition is met. This approach efficiently executes the optimized plan when no anomalies are detected and precisely activates the corresponding fault-tolerant mechanism based on the root cause type when an anomaly is detected, thereby providing high robustness for temperature maintenance control of lithium titanate power batteries in low-temperature environments.

[0091] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0092] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0093] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wireless or wired transmission; wired transmission methods include optical fiber, twisted pair, coaxial cable, etc.; wireless transmission includes infrared, microwave, etc. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center containing one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0094] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0095] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0096] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0097] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0098] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0099] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0100] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for maintaining the temperature of a lithium titanate power battery suitable for low-temperature environments, characterized in that, include: S1. Obtain battery temperature data collected by the vehicle battery management system and heater status data collected by the charging pile control system; S2. When the battery temperature data is lower than the preset preheating start-up temperature threshold, a time-series analysis is performed on the battery temperature data and heater status data based on the historical collaborative operation data of the vehicle battery management system and the charging pile control system to identify causal relationship patterns, including: Extract time-series data segments with the same operating conditions as the current battery temperature data and heater status data from the historical collaborative operation data of the vehicle battery management system and the charging pile control system. The battery temperature data sequence and heater status data sequence in the extracted time-series data segments are time-aligned to form a synchronized, collaborative operation data pair sequence; The dynamic correlation strength between battery temperature data and heater status data in the synchronized collaborative operation data sequence under multiple time delays is quantified. Based on the distribution and statistical significance of dynamic correlation strength under multiple time delays, a stable causal relationship from heater state data to battery temperature data is identified as a causal correlation pattern. S3. Based on the causal relationship model, predict the collaborative failure risk parameters corresponding to different preheating control allocation schemes; the preheating control allocation scheme defines the authority, timing and logical rules for the on-board battery management system and the charging pile control system to issue control commands to the heater during the preheating process. S4. Based on the causal relationship model, conduct a causal counterfactual vulnerability assessment of the preheating control allocation scheme with the lowest collaborative failure risk parameter. S5. Based on the results of the collaborative failure risk parameters and vulnerability assessment, select the final execution plan, and perform trend analysis on the real-time acquired battery temperature data and heater status data to determine potential abnormal signs and their source types. S6. If there are no potential abnormal signs, execute the final execution plan; if there are potential abnormal signs, select the corresponding fault-tolerant control strategy according to the source type and execute the heating control.

2. The method for maintaining the temperature of a lithium titanate power battery suitable for low-temperature environments according to claim 1, characterized in that, S1 includes: Through the communication interface between the vehicle battery management system and the charging pile control system, the battery temperature data periodically sent by the vehicle battery management system is received in real time. Through the communication interface, the system receives real-time heater status data, including heater start / stop status and operating power, sent by the charging pile control system during heater operation.

3. The method for maintaining the temperature of a lithium titanate power battery suitable for low-temperature environments according to claim 1, characterized in that, S3 includes: Based on the causal influence strength and time delay from heater state data to battery temperature data in the causal association model, the causal chain maintenance conditions required to maintain effective heating are determined. For each preheating control allocation scheme, analyze the control logic of the preheating control allocation scheme on the heater status data, and determine whether the control logic satisfies the causal chain maintenance condition. The probability of a break in the causal relationship between heater state data and battery temperature data due to failure to meet the causal chain maintenance condition is quantified and used as a collaborative failure risk parameter corresponding to the preheating control allocation scheme.

4. The method for maintaining the temperature of a lithium titanate power battery suitable for low-temperature environments according to claim 1, characterized in that, S4 includes: Based on the causal association model, at least one counterfactual condition is constructed that can disrupt the causal chain from heater state data to battery temperature data. Under the premise of keeping other causal relationships unchanged in the causal relationship model, we infer the control effect of the preheating control allocation scheme with the lowest collaborative failure risk parameter on battery temperature data when counterfactual conditions occur. Based on the degree of deviation between the simulated control effect and the expected control effect without counterfactual conditions, assess the vulnerability of the preheating control allocation scheme under abnormal conditions.

5. A method for maintaining the temperature of a lithium titanate power battery suitable for low-temperature environments according to claim 4, characterized in that, Based on the degree of deviation between the simulated control effect and the expected control effect without counterfactual conditions, assess the vulnerability of the preheating control allocation scheme under abnormal conditions, including: The control effect of the battery temperature data obtained under the counterfactual conditions is compared with the expected control effect under normal conditions based on the causal relationship model. Calculate the difference between the two in terms of key performance indicators, where the key performance indicators include at least one of the decrease in temperature maintenance stability and the time delay in achieving the target temperature; Based on the comparison between the degree of difference and the preset degree of difference threshold, the vulnerability level of the preheating control allocation scheme under abnormal conditions is determined.

6. A method for maintaining the temperature of a lithium titanate power battery suitable for low-temperature environments according to claim 1, characterized in that, S5 include: Based on the complementary relationship between the collaborative failure risk parameters and the vulnerability assessment results, a scheme selection rule is constructed, and the final execution scheme is selected from the schemes with the lowest collaborative failure risk parameters based on the scheme selection rule; Extract the trend characteristics of change within the current time window from the real-time acquired battery temperature data and heater status data; The matching degree is calculated between the changing trend characteristics of battery temperature data and the changing trend characteristics of heater status data to identify whether the changing trend characteristics of the two deviate from the normal cooperative pattern established by the causal relationship pattern. When a deviation is identified, the deviation characteristics are compared with a pre-defined library of typical abnormal patterns to determine the source type of the potential abnormality.

7. A method for maintaining the temperature of a lithium titanate power battery suitable for low-temperature environments according to claim 6, characterized in that, Source types include those indicated by abnormal battery temperature data, those indicated by abnormal heater status data, and those indicated by abnormal interaction between the two.

8. A method for maintaining the temperature of a lithium titanate power battery suitable for low-temperature environments according to claim 1, characterized in that, S6 include: After identifying potential anomalies, the target fault-tolerant control strategy corresponding to the source type is matched from the preset set of fault-tolerant control strategies based on the determined source type of the potential anomalies. When the source type is indicated by abnormal battery temperature data, the first type of fault-tolerant control strategy is executed; When the source type is indicated by the abnormality of heater status data, the second type of fault-tolerant control strategy is executed; When the source type is the type indicated by the abnormal relationship between the two, the third type of fault tolerance control strategy is executed.

9. A method for maintaining the temperature of a lithium titanate power battery suitable for low-temperature environments according to claim 8, characterized in that, The first type of fault-tolerant control strategy includes increasing the sampling frequency of battery temperature data and dynamically adjusting the control commands for heater status data based on the latest sampling data; the second type of fault-tolerant control strategy includes switching to a backup communication link to obtain heater status data and adjusting the control commands based on the verification results of redundant data; the third type of fault-tolerant control strategy includes activating a third-party temperature monitoring unit independent of the vehicle battery management system and the charging pile control system, and executing heating control based on the verification data of the third-party temperature monitoring unit.

Images